The Azores Bullfinch

bs_bs_banner

Biological Journal of the Linnean Society, 2013, 108, 677–687. With 4 ﬁgures

The Azores bullﬁnch (Pyrrhula murina) has the same unusual and size-variable sperm morphology as the Eurasian bullﬁnch (Pyrrhula pyrrhula)

JAN T. LIFJELD1*, ANTJE HOENEN2, LARS ERIK JOHANNESSEN1, TERJE LASKEMOEN1, RICARDO J. LOPES3, PEDRO RODRIGUES3,4 and MELISSAH ROWE1

1Natural History Museum, University of Oslo, PO Box 1172 Blindern, 0318 Oslo, Norway 2Electron Microscopical Unit for Biological Sciences, Department of Molecular Biosciences, University of Oslo, PO Box 1041 Blindern, 0316 Oslo, Norway 3CIBIO, Centro de Investigação em Biodiversidade e Recursos Genéticos, InBIO Laboratório Associado, Universidade do Porto, 4485-661 Vairão, Portugal 4CIBIO, Centro de Investigação em Biodiversidade e Recursos Genéticos, InBIO Laboratório Associado, Polo dos Açores, Universidade dos Açores, 9501-801 Ponta Delgada, Portugal

Received 25 July 2012; revised 25 September 2012; accepted for publication 25 September 2012

The Azores bullfinch is endemic to the island of São Miguel in the Azores archipelago and the sister species to the Eurasian bullfinch. Here we show that the spermatozoa of the two species have similar ultrastructure and gross morphology. Thus, the unusual and supposedly neotenous sperm morphology previously described for the Eurasian bullfinch appears to be an ancestral trait that evolved before the two taxa diverged. In addition, the coefficients of variation in total sperm length, both within and among males, were high in both species and exceed any previously published values for free-living passerines. Such high sperm-size variation is typically found in species with relaxed sperm competition. However, the high variance in mean sperm length among Azores bullfinches is surprising, because the trait has high heritability and this small, insular population shows clear signs of reduced genetic diversity at neutral loci. A possible explanation for this apparent contradiction is that the Azores bullfinch has retained more diversity at functional and fitness-related loci than at more neutral parts of the genome. Finally, we also present data on relative testis size and sperm swimming speed for the Eurasian bullfinch, and discuss the hypothesis that the small and putatively neotenous sperm in bullfinches has evolved in response to lack of sperm competition. © 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 108, 677–687.

ADDITIONAL KEYWORDS: Fringillidae – Passeriformes – sperm competition – sperm length – sperm velocity – testis size.

INTRODUCTION the order Passeriformes has a unique and highly conserved sperm structure: sperm are filiform, the Sperm cells are extremely variable in shape and size head (i.e. acrosome and nucleus) is corkscrew-shaped, across the animal kingdom (Cohen, 1977; Pitnick, the midpiece is elongated and wrapped around the Hosken & Birkhead, 2009). Phylogeny explains much flagellum, and a helical membrane (microtubule of this variation, which makes sperm morphology an helix) is coiled along the length of the head and informative trait for phylogenetic inference and tax- midpiece (Koehler, 1995). This is in contrast to onomy (Jamieson, Ausió & Justine, 1995). In birds, the non-passeriform orders, in which sperm cells, while also filiform, typically have a smooth and cylindrical head, a short midpiece, and no helical mem- *Corresponding author. E-mail: [email protected] brane (Jamieson, 2006). There are also a number

© 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 108, 677–687 677 678 J. T. LIFJELD ET AL. of ultrastructural details visible through scanning (2) the Himalayan clade with the orange bullfinch (SEM) and transmission electron microscopy (TEM) (P. auranthiaca), the red-headed bullfinch (P. erythro- that make the sperm of oscine passerines (Passeri) cephala), and the grey-headed (or Beavan’s) bullfinch unique among birds, including a single, fused mito- (P. erythaca); and (3) the Southeast Asian clade with chondrion, the lack of a proximal centriole (but see the brown bullfinch (P. nipalensis) and the white- Aire & Ozegbe, 2012) and an elongated, pointed acro- cheeked bullfinch (P. leucogenis) (Töpfer et al., 2011). some (Koehler, 1995; Jamieson, 2006). Importantly, The last-named clade is more basal in the bullfinch passerine sperm vary greatly in size and dimensions phylogeny, leaving the Himalayan clade as the sister (Koehler, 1995; Jamieson, 2006; Immler & Birkhead, to the Eurasian clade. To date, sperm morphology 2007; Kleven et al., 2008): the mean total length of has not been described in these species, with the formalin-fixed sperm ranges between 40 and 280 mm exception of one individual of the grey-headed bull- among the more than 200 passerine species screened finch (see Birkhead et al., 2006). In that instance, so far in our lab (Lifjeld et al., 2010; own unpublished sperm were found to have a pointed and spiral- data). Most of this variation is attributable to the shaped head (i.e. the typical passerine head shape length of the midpiece (Immler & Birkhead, 2007), contrasting markedly with that of the Eurasian bull- which contains the fused mitochondria important for finch sperm), although sperm length was equally energy production and sperm movement (Cummins, short with an extremely small midpiece. This finding 2009). would suggest that the neotenous sperm of the Eura- One species stands out as an exception to the sian bullfinch might be an autapomorphy (i.e. con- typical passerine sperm, namely the Eurasian bull- fined to the Eurasian bullfinch only). However, more finch Pyrrhula pyrrhula (Birkhead et al., 2006, 2007). species in this genus need to be analysed before any In this species, spermatozoa are non-filiform and definitive conclusions on the evolution of the unusual have a cylindrical head with a rounded, not pointed sperm morphology in the Eurasian bullfinch can be acrosome and a very short midpiece that lacks a made. helical membrane (Birkhead et al., 2006). In addition, The Azores bullfinch is endemic to the island São ultrastructural studies reveal several primitive traits Miguel in the Azores archipelago in the Atlantic not found in other oscine passerines, including a Ocean (BirdLife International, 2010; del Hoyo et al., proximal centriole and several small mitochondria 2010). The breeding range of this species is restricted instead of one large, fused mitochondrion (Birkhead to a 100-km2 area of native laurel forest on the et al., 2007). However, the deviant structural charac- eastern side of the island, and the current population teristics of the sperm of the Eurasian bullfinch are size is estimated at about 1000–1600 individuals actually present in the spermatids (i.e. immature (Monticelli et al., 2010; Ceia et al., 2011). The species sperm) of other passerines. Thus, suppression of the has recently been downlisted from ‘Critically Endan- final stages of spermatogenesis is thought to cause gered’ to ‘Endangered’ on the IUCN Red List, as the the unusual morphology of the Eurasian bullfinch population size is considered to be stable (BirdLife sperm (Birkhead et al., 2007). It has been further International, 2010). However, the species underwent hypothesized that the neotenous sperm of this species a severe population bottleneck in the 20th century, may ultimately result from a lack of sperm competi- and the current population shows clear signs of low tion (Birkhead et al., 2006; Birkhead & Immler, 2007). genetic diversity. There seems to be only one mito- Several lines of indirect evidence, including low rela- chondrial haplotype in the current population (Töpfer tive testis mass, few sperm stored in the seminal et al., 2011; R. J. Lopes, unpubl. data). A microsatel- glomera and low sperm swimming speed (Birkhead lite analysis also revealed very low levels of hetero- et al., 2006), as well as relatively short sperm with zygosity and allelic diversity relative to continental high coefficients of intermale variation in sperm total populations of the Eurasian bullfinch (R. J. Lopes, length and size components (Calhim, Immler & Birk- unpubl. data). head, 2007), support this assumption. However, to Here we present the first description of the sperm date there are no paternity studies published on the morphology and its variation in the Azores bullfinch. species to confirm genetic monogamy. Our study has two main objectives. First, we The evolutionary origin of the unusual sperm mor- wanted to know whether the neotenous sperm morphology of the Eurasian bullfinch is poorly under- phology of the Eurasian bullfinch is also present in stood. According to the most recent classifications the Azores bullfinch. If so, the trait might be a (del Hoyo, Elliott & Christie, 2010; Sangster et al., synapomorphy present in the last common ancestor 2011; Gill & Donsker, 2012), the genus Pyrrhula of the entire Eurasian bullfinch clade, as opposed to consists of seven species. These species fall into three a more recent event within the Eurasian bullfinch major phylogenetic clades: (1) the Eurasian clade lineage. We studied the sperm morphology of both with P. pyrrhula and the Azores bullfinch (P. murina); species using SEM and TEM for ultrastructural

© 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 108, 677–687 SPERM MORPHOLOGY IN BULLFINCHES 679 details, and high-resolution light microscopy for MATERIAL AND METHODS gross morphology. Second, we wanted to examine if the Azores bull- The analyses of sperm morphology include samples finch and the Eurasian bullfinch differ in sperm mor- from 11 male Azores bullfinches and 13 Eurasian phometry: sperm size and inter- and intramale bullfinches. The Azores bullfinches were caught in variation in sperm length. Sperm length has a strong mist-nets erected at foraging sites throughout the genetic basis (Birkhead et al., 2005; Simmons & distribution range, in São Miguel island, during 13–17 Moore, 2009). Thus, high phenotypic variance in June 2011. The Eurasian bullfinches were caught in sperm morphology among males in a population indi- mist-nets at feeding trays baited with sunflower seeds cates high additive genetic variance for the trait. at various locations in and around Oslo, south-east Given the low population genetic diversity in the Norway, in April–June 2006–2012. The technique of Azores bullfinch (R. J. Lopes, unpubl. data), one sperm sampling followed the protocol of cloacal would expect reduced phenotypic variance in sperm massage: collection of an ejaculate in a capillary tube, length among males in the population. Knowledge of and subsequent dilution in phosphate-buffered saline intermale variation in sperm length also provides before fixation in 5% formalin (Kleven et al., 2008). insight into sperm competition levels faced by males Four of the Eurasian bullfinches (all from 2012) were within a species. Specifically, there is robust compara- killed under licence for another study, and their testes tive evidence that the coefficient of variation in sperm dissected out and weighed (± 0.001 g). All other birds length among males in a population decreases with were released unharmed after sampling. increasing risks of sperm competition (Calhim et al., For SEM, sperm were attached to glass coverslips 2007; Kleven et al., 2008; Lifjeld et al., 2010). An precoated with poly-lysine (1 mg mL-1; Sigma P1274) interpretation of this association is that sperm com- by placing a 10-mL aliquot of formalin-fixed sperm petition acts as a stabilizing selection pressure on onto each coverslip and incubating samples overnight sperm length. Thus, both reduced genetic diversity in a wet chamber at room temperature. Next, sperm and increased sperm competition are plausible and adhering to the coverslips were dehydrated using a alternative explanations for reduced intermale vari- graded ethanol series (5, 10, 15, 20, 30, 40, 50, 60, 70, ance in sperm length, and we examine this variation 80, 90, 95 and 100% ethanol) and critical point dried in order to discuss hypotheses of sperm evolution in (BAL-TEC CPD 030 Critical Point Dryer). Coverslips bullfinches. were then mounted on SEM stubs using carbon tape Similarly, we compared the levels of intramale vari- and sputter coated with 4–6 nm platinum using a ation in sperm length. Heterogeneous sperm within Cressington 308R coating system. Finally, samples an ejaculate is thought to reflect developmental insta- were examined and digital images recorded using bility or imperfections in the sperm production a Hitachi S-4800 field emission scanning electron machinery associated with relaxed sperm competition microscope operated at 5.0 kV. (Immler, Calhim & Birkhead, 2008; Kleven et al., For TEM (Azores bullfinch only), formalin-fixed 2008), inbreeding or low genetic diversity (Gage et al., sperm were further fixed in 2.5% glutaraldehyde in 2006; Roldan & Gomendio, 2009), or environmental 0.1 M sodium cacodylate buffer (pH 7.4) for 1 h at stressors or pollutants (Pitnick et al., 2009). In con- room temperature (RT), and then post-fixed in 2% trast to the Eurasian bullfinch, which exhibits no osmium tetraoxide in 0.1 M cacodylate buffer in the lower genetic diversity than other European species in dark, for 1 h at RT. Samples were then processed for the family Fringillidae (Durrant et al., 2010), the embedding in a Pelco BioWave following a multi-step Azores bullfinch has significantly reduced genetic protocol. First samples were processed in triplicate in diversity (R. J. Lopes, unpubl. data). Consequently, we 1% uranyl acetate in double distilled water (3 min at predicted that the Azores bullfinch should exhibit high 37 °C/250 W; 3 min at RT; 3 min at 37 °C/250 W). levels of intramale variation in sperm morphology. Next sperm were rinsed three times in distilled water Finally, we also present some quantitative data on (each rinse for 60 s at 37 °C/250 W), and dehydrated testis size and sperm swimming speed in the Eura- using a graded acetone series (50, 70, 90, 100, 100 sian bullfinch for comparison with published data and 100%; each step for 60 s at 37 °C/250 W). Sperm from this and other species. Typically, bird species were then infiltrated with resin at increasing concen- with low sperm competition have relatively small trations (25, 50 and 75%, each for 3 min at 45 °C/ testes (Møller & Briskie, 1995) and slow-swimming 350 W; followed by 100, 100 and 100% each for 3 min sperm (Kleven et al., 2009). Thus, examination of at 50 °C/350 W). Between each step in this process these traits will help provide support for or against samples were centrifuged at 600 g for 15 min. the hypothesis that the unusual sperm morphology of Next, samples were polymerized overnight at 60 °C. Eurasian bullfinch is due to an absence of sperm Finally, sections (80 nm) were obtained with the competition. Leica microtome, post-stained with lead citrate and

© 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 108, 677–687 680 J. T. LIFJELD ET AL. observed using a Philips CM200 transmission elec- ment (ALH) and beat cross frequency (BCF). Mean tron microscope operated at 80 kV. motility measurements were calculated from all For sperm morphometrics, a small aliquot of motile spermatozoa to obtain one overall measure- approximately 15 mL of the formalin–sperm solution ment for each ejaculate. Spermatozoa with was applied onto a microscope slide and was allowed VAP < 30 mms-1 and/or VSL < 25 mms-1 were consid- to air-dry before inspection, digital imaging, and ered static or drifting and excluded from the motility measurement under light microscopy. We took digital analyses, along with spermatozoa tracked for less images of spermatozoa at magnifications of 640¥, than ten frames. We also removed any tracks for using a Leica DFC420 camera mounted on a Leica which the maximum frame-to-frame movement DM6000 B digital light microscope. The morphomet- exceeded the average frame-to-frame movement by 4 ric measurements were conducted using Leica Appli- SDs for the same track, as such tracks tended to cation Suite (version 2.6.0 R1). We measured the represent tracking errors in the software. Because the length of head (including the acrosome, nucleus, and spermatozoa appeared to have size-variable, circular midpiece) and the flagellum (± 0.1 mm) of ten intact heads (see example video in Supporting Information), spermatozoa per male. Total sperm length was calcu- it was not possible to distinguish dead/immotile sper- lated as the sum of head and flagellum length. We matozoa from faeces debris in the sample. Conse- have previously shown that sperm length measure- quently, we did not estimate the percentage motile ments have very low measurement error and high sperm in the samples. For logistical reasons, sperm repeatability (Laskemoen et al., 2007, 2010). Stand- velocity recordings could not be made for the Azores ardized values of intra- and intermale variation was bullfinch. Raw data for the analyses of sperm mor- expressed as the coefficient of variation (CV = SD/ phometrics and sperm velocity can be downloaded mean ¥ 100). For intermale variation, a correction from the Dryad database (Lifjeld et al. 2012). factor for variation in sample size (N) was applied, namely CV = SD/mean ¥ 100 ¥ (1 + 1/4N), as recom- RESULTS mended by Sokal & Rohlf (1981). For sperm velocity recordings, we collected fresh The scanning electron micrographs revealed that the ejaculates from 11 Eurasian bullfinches and immedi- Azores bullfinch and the Eurasian bullfinch have the ately diluted them in 30 mL of preheated Dulbecco’s same sperm morphology (Figs 1, 2). Both have non- modified Eagle medium (advanced DMEM; Invitro- filiform sperm showing a cylindrical, thickened, and gen, Carlsbad, CA, USA) and placed the tubes in a tapering head with a rounded acrosome as a cap on heating block set to 40 °C. Next, we carefully homog- top of the nucleus, an extremely small midpiece, and enized the solution using a pipette and within 20 s no helical membrane wound around the flagellum. 5–6 mL of the diluted sperm was deposited on a There was, however, much variation in head elonga- preheated counting chamber (20 mm depth, two- tion, even among sperm from the same ejaculate (e.g. chamber slide, Leja Products BV, Nieuw-Vennep, the Fig. 2), and for both species the length of the acro- Netherlands), mounted on a Hamilton Thorne some was generally shorter than the nucleus. There MiniTherm stage warmer set to 40 °C (Hamilton was also considerable variation in the surface texture Thorne Biosciences, Beverly, MA, USA). Sperm of the head. In general, the acrosome had a smoother motion was recorded using a mini-DV camera (Sony surface than the nucleus, which mostly had a very HDR-HC1E or Canon Legria HF S200) mounted on rugged and furrowed appearance. However, we an upright microscope (Olympus CX41) with a 4¥ cannot exclude the possibility that the surface struc- objective. The camera was set to 10¥ optical zoom and ture has been altered during sample fixation (see infinite focus. For each male we recorded six different below). The microtubular helix along the flagellum, viewing fields of the counting chamber. Computer- being visible in both species (Figs 1, 2), reveals the assisted sperm analysis (HTM-CEROS sperm tracker, only obvious diagnostic passerine trait of the external CEROS version 12, Hamilton Thorne Research) was morphology of the spermatozoa. used to analyse the recordings. The sperm analyser Transmission electron micrographs of the Azores was set at a frame rate of 50 Hz and 25 frames (i.e. bullfinch spermatozoon confirmed the presence of sperm cells were tracked for 0.5 s). Each analysis was multiple mitochondria in the midpiece (Fig. 3). There visually examined and cell detection parameters were was also marked variation in the density of chromatin adjusted to optimize the detection of motile sperma- within the nucleus. The outer membrane of the tozoa. Cell detection parameters thus varied slightly; nucleus and midpiece appeared broken in places minimum contrast 60–140, and minimum cell detec- (Fig. 3A–C), suggesting that the use of formalin-fixed tion size 10–12 pixels. We recorded average path samples may not be optimal for electron microscopy of velocity (VAP), straight line velocity (VSL), curvilin- these delicate structures. We were also not able to ear velocity (VCL), amplitude of lateral head displace- identify any centrioles, and thus we could not confirm

Figure 1. Scanning electron micrographs of the head and anterior part of the flagellum of the Azores bullfinch spermatozoon. The main structural parts are indicated with their names, and the arrow points to the microtubular helix on the flagellum.

Figure 2. Scanning electron micrographs of the head and anterior part of the flagellum of the Eurasian bullfinch spermatozoon. The main structural parts are indicated by their names and the microtubular helix on the flagellum by an arrow. Note the variation in head shape among individual spermatozoa (indicated by *). the presence of a proximal centriole in the Azores variation in mean sperm length among males in the bullfinch spermatozoon. two species, with a size range of 39–52 mm for the Sperm size dimensions were also similar for the two 11 Azores bullfinches and 40–53 mm for the 13 species (Table 1). Total sperm length averaged about Eurasian bullfinches (Fig. 4). The coefficient of inter-

46 mm in both, which is very similar to the value of male variation in mean sperm length (CVbm) was 9.6 46.87 mm reported by Birkhead et al. (2006) for the and 8.6% (Azores bullfinch and Eurasian bullfinch, Eurasian bullfinch. There was also a considerable respectively), and the two variances did not differ

Figure 3. Ultrastructure of the spermatozoon of the Azores bullﬁnch by transmission electron microscopy (TEM). A, longitudinal section (LS) of a spermatozoon showing the midpiece consisting of multiple mitochondria grouped at the base of the nucleus. B, LS section of a spermatozoon showing the cap-like acrosome (acrosomal vesicle). C, transverse section (TS) of a spermatozoon showing several discrete mitochondria and the nucleus, in which the chromatin is loosely arranged. D, TS of the ﬂagellum showing the archetypical 9 + 2 microtubule structure of the axoneme (two central microtubule singlets and nine outer microtubule doublets).

Table 1. Sperm morphometrics in the Azores bullﬁnch and the Eurasian bullﬁnch based on the measurement of ten sperm cells per male; all measurements are in mm, and mean values are given with ± SD

Azores bullﬁnch Eurasian bullﬁnch Test of difference Test of difference Sperm trait P. murina (N = 11) P. pyrrhula (N = 13) between means between variances

Head length* 5.60 ± 0.29 5.78 ± 0.37 t =-1.35, P = 0.19 F-ratio = 1.67, P = 0.42 Flagellum length 39.99 ± 4.23 40.49 ± 4.30 t =-0.31, P = 0.76 F-ratio = 1.20, P = 0.75 Total length 45.58 ± 4.29 46.28 ± 4.31 t =-0.41, P = 0.68 F-ratio = 1.22, P = 0.73

CVwm of total length 6.07 ± 2.28 7.51 ± 1.80 t =-1.72, P = 0.10 F-ratio = 1.60, P = 0.43

CVbm of mean total length† 9.62 8.62

*Acrosome, nucleus, and midpiece.

†Adjusted for low sample sizes by the formula CVbm = SD/mean ¥ 100 ¥ (1 + 1/4N) (Sokal & Rohlf, 1981). significantly (Table 1). Both values exceeded the (Lifjeld et al., 2010). Finally, neither head length nor highest value of CVbm (= 6.2%) recorded among 55 flagellum length differed significantly between the passerine species reported by Lifjeld et al. (2010). The two bullfinches, nor did their variances (Table 1). coefficients of intramale variation in total sperm For the four collected Eurasian bullfinches, the length were also extremely high (mean CVwm of 6.1 mass of the left and the right testis averaged and 7.7%, Table 1), and exceeded the maximum value 0.041 ± 0.016 g (mean ± SD) and 0.025 ± 0.006 g, (= 3.6%) observed among 55 other passerine species respectively. With a mean body mass of 29.9 g for

Figure 4. Variation in total sperm length among and within males of the Azores bullﬁnch and Eurasian bullﬁnch. Ten spermatozoa were measured for each male, and the mean and SD of total sperm length are indicated by squares and vertical bars, respectively. Within each species, individual males are ordered by increasing mean values.

Table 2. Sperm motility measurements in the Eurasian to date in passerines; mean VCL estimates published bullfinch from video recordings of semen samples of wild- for 42 passerine species varied from 77 to 164 mms-1 caught birds (N = 11) at standard conditions of 40 °C, and using the same methodology (Kleven et al., 2009). The using Computer-Assisted Sperm Analysis (CASA) with a three velocity parameters VCL, VAP and VSL were filtered output (see Methods for details) highly intercorrelated (Pearson’s r > 0.89, P < 0.001), whereas the ALH and BCF parameters were not Parameter Measurement significantly correlated with any other motility measures. The ALH and BCF estimates lie within Velocity curvilinear (VCL) 157.3 ± 22.8 (127.3–202.3) the normal range for passerines (J. T. Lifjeld et al., (mms-1) unpubl. data). A video clip of swimming sperm of Velocity average path (VAP) 150.4 ± 21.0 (122.2–185.5) the Eurasian bullfinch is available as Supporting (mms-1) Information. Velocity straight line (VSL) 144.8 ± 19.2 (119.8–169.7) (mms-1) Amplitude of lateral head 3.2 ± 1.7 (1.4–7.0) displacement (ALH) (mm) DISCUSSION Beat cross frequency (BCF) 21.6 ± 3.2 (14.9–25.8) PHYLOGENY OF THE UNUSUAL (Hz) BULLFINCH SPERMATOZOA

Measurements are given as mean ± SD (range) and were Our study showed that the spermatozoa of the calculated from an average of 97 ± 113 SD sperm tracks Azores bullfinch and the Eurasian bullfinch have the per individual (range 8–353). same gross morphology and ultrastructure. The simi- larity is not only restricted to the unusual form and shape of the head and flagellum, as previously these males, the combined relative testis mass described for the Eurasian bullfinch (Birkhead et al., amounted to only 0.22%. This corresponds well with 2006, 2007), but there were also no differences in the value of 0.29% reported by Birkhead et al. (2006) sperm morphometrics between the two species. Both for the same species, and implies extremely small species exhibit an extremely high variation in sperm testes for its body size compared with most other size, both among individual sperm within an ejacu- passerine species (Møller, 1991; Birkhead et al., late and in mean sperm size among males. Hence, 2006). the two species seem indistinguishable based on Values of sperm swimming speed for the 11 Eura- their spermatozoa. sian bullfinches were surprisingly high (Table 2): The lack of differentiation between the two sister mean curvilinear velocity (VCL) was 157 mms-1. taxa suggests an ancestral origin of this exceptional Importantly, this value is among the highest records sperm morphology, as opposed to a more recent or

© 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 108, 677–687 684 J. T. LIFJELD ET AL. contemporary event in the Eurasian bullfinch lineage. et al., 2010). However, in contrast to expectations, we The Eurasian bullfinch shows large geographical vari- found that the sperm of the Eurasian bullfinch swim ation in plumage morphology, and 11 subspecies are at very high speeds (157 mms-1), which is not only currently recognized within its wide Palearctic range high compared with other passerine species (Kleven (Gill & Donsker, 2012). If it is correct that the Azores et al., 2009), but also with a previous report of only 22 bullfinch is basal in the phylogeny of the Eurasian mms-1 in two captive-bred Eurasian bullfinches (Birk- clade (Töpfer et al., 2011), then all Eurasian bullfinch head et al., 2006). However, the latter study also populations and subspecies should share the same reported similarly low velocities for ten other passer- unusual sperm morphology. This prediction obviously ine species (range 21–43 mms-1), and thus results do warrants further empirical studies, especially of the not appear comparable between the two studies, pre- eastern subspecies griseiventris and cineracea, both of sumably due to different methodologies. Importantly, which are morphologically and genetically distinct although the high sperm velocity in our Eurasian (del Hoyo et al., 2010; Töpfer et al., 2011). The Hima- bullfinches might be taken as an indicator of high layan clade, with three species, is sister to the Eura- sperm competition risk from a comparative perspec- sian clade (Töpfer et al., 2011), and one of the species, tive (cf. Kleven et al., 2009), we would caution against the Grey-headed bullfinch, has sperm with a helical such an interpretation due to the deviant morphology head shape (Birkhead et al., 2006). This would of the bullfinch sperm. Specifically, the ways in which suggest that the entire clade has retained the typical variation in sperm morphology translates into varia- passerine sperm structure, and thus that the unusual tion in sperm velocity in passerine birds is poorly bullfinch sperm evolved early in the Eurasian lineage understood. Moreover, it should be noted that, under after the split from the Himalayan clade. Nonethe- in vitro conditions, sperm swim in an artificial less, more information about sperm morphology of medium that may not accurately mimic the female all Pyrrhula species is needed to resolve this issue, environment. We therefore agree with Birkhead et al. and also to test whether there is an abrupt shift or (2006) that the current evidence seems to indicate a gradual change in sperm morphology across the that sperm competition is low or absent in the Eura- Pyrrhula phylogeny. sian bullfinch, and we suggest that it is similarly low or absent in the sister species, the Azores bullfinch.

SPERM COMPETITION IN BULLFINCHES In both bullfinch species, the spermatozoa are very GENETIC DIVERSITY short. A mean value of 46 mm is towards the lower end Durrant et al. (2010) suggested that the unusual of the length distribution for passerine birds. For sperm morphology in the Eurasian bullfinch is an example, total sperm length varies between 43 and inbreeding effect caused by a reduction in genetic 280 mm in a comparative study of 55 passerine species diversity. Abnormal sperm is typically more frequent from 21 families (Lifjeld et al., 2010). Interestingly, in endangered or captive populations with reduced the Fringillidae family spans almost the same range, heterozygosity (Gage et al., 2006; Fitzpatrick & from 46 mm in the Eurasian and Azores bullfinch (this Evans, 2009; Roldan & Gomendio, 2009) and the study) to 275 mm in the common rosefinch (Carpoda- Eurasian bullfinch may have undergone several popu- cus erythrinus; Lifjeld et al., 2010). This implies that lation bottlenecks (Durrant et al., 2010). Deprivation there is also considerable variation in sperm compe- of genetic diversity is also documented in the endan- tition risk within the family, as a positive association gered Azores bullfinch (R. J. Lopes, unpubl. data). between sperm length and sperm competition risk However, the analysis of microsatellite markers in the has been documented for passerine birds (Briskie, Eurasian bullfinch did not reveal any signature of a Montgomerie & Birkhead, 1997; Kleven et al., 2009; recent bottleneck or signs of reduced genetic variation Lüpold, Linz & Birkhead, 2009), and for the Fringil- when compared with three other fringillid finches lidae family (Immler & Birkhead, 2007). The short (Durrant et al., 2010). Likewise, microsatellite allelic spermatozoa in the bullfinches are thus indicative of richness and heterozygosity levels are higher in the low sperm competition risk in these species. Eurasian bullfinch than in the Azores bullfinch (R. J. Further evidence for low sperm competition in bull- Lopes, unpubl. data). Therefore, it seems unlikely finches comes from the very small relative testes size that the unusual sperm morphology in the two bull- (Birkhead et al., 2006; this study), the small cloacal finch species can be explained by inbreeding effects. protuberance (Birkhead et al., 2006; own personal We would argue, however, that considering the observations) and the high coefficients (CV) of both typical bullfinch sperm as ‘abnormal’ is inappropriate. intra-male and inter-male variation in sperm total Rather, the term should be used to describe sperm length compared with other passerines (Calhim et al., that deviates markedly in shape or structure from the 2007; Immler et al., 2008; Kleven et al., 2008; Lifjeld normal or average sperm phenotype within a given

© 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 108, 677–687 SPERM MORPHOLOGY IN BULLFINCHES 685 species. In the two bullfinch species, the non-filiform from a suppression of the final stages of spermatogen- sperm phenotype is universal and we saw no deviant esis (Birkhead et al., 2007). A similar phenomenon of types; thus for these species, a non-filiform, rounded evolution of simplified sperm is also known from sperm is normal and does not represent an abnormal other vertebrates. For example, in muroid rodents, sperm morphology. Moreover, the high intraspecific several species have lost the ancestral apical hook variation in sperm size and dimensions seen in bull- of the sperm head and evolved a rounded acrosome, finches reflects a common phenomenon seen in every poor condensation of chromatin in the nucleus, and species, that sperm is variable both within and among a shorter flagellum (Breed, 2005). Typically, these males (Birkhead & Immler, 2007; Pitnick et al., 2009). species have small relative testis size and high What makes the bullfinches unique in this respect is intramale variability in sperm components, which that this normal variation is higher than in most imply relaxed sperm competition (Breed, 2005; Breed other species, a result which is also consistent with et al., 2007). Similarly, in naked mole rats (Hetero- the hypothesis of relaxed sperm competition in these cephalus glaber), the sperm is simplified with several species. putatively neotenous features, and the mating system That the Azores bullfinch is deprived of genetic is characterized by low sperm competition (van der diversity while still maintaining high phenotypic Horst et al., 2011). It therefore seems to be a common diversity in sperm size, a genetically determined trait pattern that atypical and simplified sperm forms (Birkhead et al., 2005; Simmons & Moore, 2009), evolve in lineages with relaxed sperm competition as seems paradoxical. One possible explanation is that an adaptation reducing the cost of sperm production the estimates of low genetic diversity primarily reflect (Wedell, Gage & Parker, 2002). We encourage more neutral genetic diversity and that functional and comparative studies of sperm morphology in verte- fitness-related loci may be less influenced by demo- brate taxa with contrasts in sperm competition levels graphic processes. Such a pattern has been reported to test the generality of this hypothesis for the evo- in the San Nicolas Island fox (Urocyon littoralis lution of simplified sperm. dickeyi), a species which has retained high genetic diversity at functional loci through balancing selection, despite a severe reduction in neutral diversity ACKNOWLEDGEMENTS (Aguilar et al., 2004). It is commonly assumed that We thank Jostein Gohli, Erica Leder, and Oddmund sperm competition exerts stabilizing selection on an Kleven for field assistance, and three anonymous optimal sperm size in passerine species (Calhim et al., referees for their helpful comments. The study was 2007; Kleven et al., 2008; Lifjeld et al., 2010), which funded by an Individual Grant (SFRH/BPD/40786/ could explain why species with intense sperm compe- 2007) from FCT (Fundação para a Ciência e Tecnolo- tition have lower intermale variation in sperm gia) to R.J.L., and a research grant (196554/V40) from size than those with little or no sperm competition. the Research Council of Norway to J.T.L. The sam- However, in species with small effective population pling was conducted under the legal permits issued by size, one would also expect the genetic variation to be the Azores Environmental Agency, Portugal, and the diminished due to drift. The fact that the Azores Directorate for Nature Management, Norway. bullfinch has maintained the same intermale variation in sperm size as the Eurasian bullfinch despite much lower population size suggests that high phe- REFERENCES notypic variance in sperm traits is somehow protected against genetic drift. A quantitative genetics study of Aguilar A, Roemer G, Debenham S, Binns M, Garcelon sperm size components in the zebra finch (Taenopygia D, Wayne RK. 2004. High MHC diversity maintained by guttata), a species which also exhibits low sperm balancing selection in an otherwise genetically monomor- phic mammal. Proceedings of the National Academy of competition (Griffith et al., 2010), revealed both sig- Sciences of the United States of America 101: 3490–3494. nificant maternal genetic effects and strong negative Aire TA, Ozegbe PC. 2012. Components and development of genetic correlations between sperm components, the centriolar complex during and beyond spermiogenesis in which could serve to constrain the effects of selection a passeridan bird, the Masked weaver (Ploceus velatus). and drift (Birkhead et al., 2005). Tissue and Cell 44: 63–67. BirdLife International. 2010. Pyrrhula murina IUCN red list of threatened species. Version 2011. 2nd edn. IUCN 2011. EVOLUTION OF NEOTENOUS SPERMATOZOA Birkhead TR, Giusti F, Immler S, Jamieson BGM. 2007. Although the bullfinch spermatozoon is unusual Ultrastructure of the unusual spermatozoon of the Eurasian among passerines, it resembles the spermatids or bullfinch (Pyrrhula pyrrhula). Acta Zoologica 88: 119–128. immature testicular passerine sperm. It has therefore Birkhead TR, Immler S. 2007. Making sperm: design, been suggested that this unusual morphology results quality control and sperm competition. In: Roldan ERS,

Gomendio M, eds. Spermatology. Nottingham: Nottingham Jamieson BGM. 2006. Avian spermatozoa: structure and University Press, 175–181. phylogeny. In: Jamieson BGM, ed. Reproductive biology and Birkhead TR, Immler S, Pellatt EJ, Freckleton R. phylogeny of aves. Enfield, NH: Science Publishers Inc., 2006. Unusual sperm morphology in the Eurasian bullfinch 249–511. (Pyrrhula pyrrhula). Auk 123: 383–392. Jamieson BGM, Ausió J, Justine J-L, eds. 1995. Advances Birkhead TR, Pellatt EJ, Brekke P, Yeates R, Castillo- in spermatozoal phylogeny and taxonomy. Paris: Muséum Juarez H. 2005. Genetic effects on sperm design in the national d’Histoire naturelle. zebra finch. Nature 434: 383–387. Kleven O, Fossøy F, Laskemoen T, Robertson RJ, Breed WG. 2005. Evolution of the spermatozoon in muroid Rudolfsen G, Lifjeld JT. 2009. Comparative evidence for rodents. Journal of Morphology 265: 271–290. the evolution of sperm swimming speed by sperm competi- Breed WG, Bauer M, Wade R, Thitipramote N, Suwa- tion and female sperm storage duration in passerine birds. jarat J, Yelland L. 2007. Intra-individual variation in Evolution 63: 2466–2473. sperm tail length in murine rodents. Journal of Zoology Kleven O, Laskemoen T, Fossøy F, Robertson RJ, Lifjeld 272: 299–304. JT. 2008. Intraspecific variation in sperm length is nega- Briskie JV, Montgomerie R, Birkhead TR. 1997. The tively related to sperm competition in passerine birds. Evo- evolution of sperm size in birds. Evolution 51: 937–945. lution 62: 494–499. Calhim S, Immler S, Birkhead TR. 2007. Postcopula- Koehler LD. 1995. Diversity of avian spermatozoa tory sexual selection is associated with reduced variation in ultrastructure with emphasis on the members of the sperm morphology. PLoS ONE 2: e413. order Passeriformes. In: Jamieson BGM, Ausió J, Justine Ceia RS, Ramos JA, Heleno RH, Hilton GM, Marques J-L, eds. Advances in spermatozoal phylogeny and tax- TA. 2011. Status assessment of the critically endangered onomy. Paris: Muséum national d’Histoire naturelle, 437– Azores bullfinch Pyrrhula murina. Bird Conservation Inter- 444. national 21: 477–489. Laskemoen T, Kleven O, Fossøy F, Lifjeld JT. 2007. Cohen J. 1977. Reproduction. London: Butterworths. Intraspecific variation in sperm length in two passerine Cummins J. 2009. Sperm motility and energetics. In: Birk- species, the bluethroat Luscinia svecica and the willow head TR, Hosken DJ, Pitnick S, eds. Sperm biology: an warbler Phylloscopus trochilus. Ornis Fennica 84: 131– evolutionary perspective. Oxford: Elsevier, 185–206. 139. Durrant KL, Dawson DA, Burke T, Birkhead TR. 2010. Laskemoen T, Kleven O, Fossøy F, Robertson RJ, The unusual sperm morphology of the Eurasian bullfinch Rudolfsen G, Lifjeld JT. 2010. Sperm quantity and (Pyrrhula pyrrhula) is not due to the phenotypic result of quality effects on fertilization success in a highly promiscu- genetic reduction. Auk 127: 832–840. ous passerine, the tree swallow Tachycineta bicolor. Behav- Fitzpatrick JL, Evans JP. 2009. Reduced heterozygosity ioral Ecology and Sociobiology 64: 1473–1483. impairs sperm quality in endangered mammals. Biology Lifjeld J, Hoenen A, Johannessen LE, Laskemoen T, Letters 5: 320–323. Lopes R, Rodrigues P, Rowe M. 2012. Data from: Gage MJG, Surridge AK, Tomkins JL, Green E, Wiskin The Azores Bullfinch (Pyrrhula murina) has the same L, Bell DJ, Hewitt GM. 2006. Reduced heterozygosity unusual and size-variable sperm morphology as the depresses sperm quality in wild rabbits, Oryctolagus cunicu- Eurasian Bullfinch (Pyrrhula pyrrhula). Dryad Digital lus. Current Biology 16: 612–617. Repository. doi:10.5061/dryad.6sf11. Gill F, Donsker D. 2012. IOC world bird names (version 3.1): Lifjeld JT, Laskemoen T, Kleven O, Albrecht T, Robert- Available at: http://www.worldbirdnames.org/ son RJ. 2010. Sperm length variation as a predictor of Griffith SC, Holleley CE, Mariette MM, Pryke SR, extrapair paternity in passerine birds. PLoS ONE 5: Svedin N. 2010. Low level of extrapair parentage in wild e13456. zebra finches. Animal Behaviour 79: 261–264. Lüpold S, Linz G, Birkhead T. 2009. Sperm design and van der Horst G, Maree L, Kotze S, O’Riain M. variation in the New World blackbirds (Icteridae). Behavio- 2011. Sperm structure and motility in the eusocial naked ral Ecology and Sociobiology 63: 899–909. mole-rat, Heterocephalus glaber: a case of degenerative Møller AP. 1991. Sperm competition, sperm depletion, orthogenesis in the absence of sperm competition? BMC paternal care, and relative testis size in birds. American Evolutionary Biology 11: 351. Naturalist 137: 882–906. del Hoyo J, Elliott A, Christie DA, eds. 2010. Handbook of Møller AP, Briskie JV. 1995. Extra-pair paternity, the birds of the world – vol. 15. Weavers to new world sperm competition and the evolution of testis size in birds. warblers. Barcelona: Lynx Edicions. Behavioral Ecology and Sociobiology 36: 357–365. Immler S, Birkhead TR. 2007. Sperm competition and Monticelli D, Ceia R, Heleno R, Laborda H, Timóteo S, sperm midpiece size: no consistent pattern in passerine Jareño D, Hilton G, Ramos J. 2010. High survival birds. Proceedings of the Royal Society B: Biological Sciences rate of a critically endangered species, the Azores bullfinch 274: 561–568. Pyrrhula murina, as a contribution to population recovery. Immler S, Calhim S, Birkhead TR. 2008. Increased Journal of Ornithology 151: 627–636. postcopulatory sexual selection reduces the intramale vari- Pitnick S, Hosken DJ, Birkhead TR. 2009. Sperm mor- ation in sperm design. Evolution 62: 1538–1543. phological diversity. In: Birkhead TR, Hosken DJ, Pitnick S,

eds. Sperm biology: an evolutionary perspective. Oxford: eds. Sperm biology: an evolutionary perspective. Oxford: Elsevier, 69–149. Elsevier, 405–434. Roldan ERS, Gomendio M. 2009. Sperm and conserva- Sokal RR, Rohlf FJ. 1981. Biometry. San Francisco: tion. In: Birkhead TR, Hosken DJ, Pitnick S, eds. Sperm W. H. Freeman and Co. biology: an evolutionary perspective. Oxford: Elsevier, 539– Töpfer T, Haring E, Birkhead TR, Lopes RJ, Severing- 564. haus LL, Martens J, Päckert M. 2011. A molecular Sangster G, Collinson JM, Crochet P-A, Knox AG, phylogeny of bullﬁnches Pyrrhula Brisson, 1760 (Aves: Parkin DT, Svensson L, Votier SC. 2011. Taxonomic Fringillidae). Molecular Phylogenetics and Evolution 58: recommendations for British birds: seventh report. Ibis 153: 271–282. 883–892. Wedell N, Gage MJG, Parker GA. 2002. Sperm competi- Simmons LW, Moore AJ. 2009. Evolutionary quantitative tion, male prudence and sperm-limited females. Trends in genetics of sperm. In: Birkhead TR, Hosken DJ, Pitnick S, Ecology and Evolution 17: 313–320.

SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article: Supplementary ﬁle. Example video of swimming sperm from an ejaculate of a Eurasian bullﬁnch as used in the CASA analysis of sperm velocity (for details see Methods).